Scan Report
48 /100
claw-shell
Run shell commands inside tmux session for AI agent interaction
Shell execution tool with declared purpose but weak safety filter that can be bypassed, allowing potentially destructive commands.
Use with caution
Add more comprehensive command validation, use allowlist approach instead of blocklist, add timeout limits, and consider sandboxing tmux session.
Findings 4 items
| Severity | Finding | Location |
|---|---|---|
| High | Unrestricted shell execution with weak safety controls RCE | handler.js:18 |
| Medium | Dangerous commands bypass filter RCE | handler.js:18 |
| Medium | Indirect environment variable access Sensitive Access | handler.js:14 |
| Low | No command timeout configured RCE | handler.js:29 |
| Resource | Declared | Inferred | Status | Evidence |
|---|---|---|---|---|
| Shell | WRITE | WRITE | ✓ Aligned | execSync used in handler.js:1,7,10 |
| Filesystem | WRITE | WRITE | ✓ Aligned | Shell can execute any file operation |
| Network | WRITE | WRITE | ✓ Aligned | Shell can execute curl/wget/netcat |
| Environment | NONE | READ | ✓ Aligned | Shell commands can read $ENV vars |
| Clipboard | NONE | NONE | — | |
| Browser | NONE | NONE | — | |
| Database | NONE | NONE | — |
1 findings
Medium External URL 外部 URL
https://x.com/... SKILL.md:39 File Tree
3 files · 2.4 KB · 106 lines JavaScript 1f · 57L
Markdown 1f · 44L
JSON 1f · 5L
├─
_meta.json
JSON
├─
handler.js
JavaScript
└─
SKILL.md
Markdown
Security Positives
✓ Purpose and capabilities are clearly documented in SKILL.md
✓ Uses dedicated tmux session 'claw' for isolation
✓ Basic safety filter exists for clearly dangerous commands
✓ No hidden functionality or obfuscation detected
✓ No credential harvesting or data exfiltration observed
✓ Single-purpose tool with focused functionality